Voltage Rise From Embedded Generators

As far as your interconnect question goes try this paper...

http://ec.europa.eu/energy/renewables/studies/doc/wind_energy/2001_fp5_brochure_energy_env.pdf

For a lot more detail about Power Systems Analysis here's the classic...

http://www.amazon.com/Power-System-Analysis-John-Grainger/dp/0070612935

You wrote "I'm quite new to 3-phase electrical systems."

This means that you should not attempt this on your own. This is a problem that seasoned engineers haven completely solved yet. There are always difficulties when connecting a power source to the grid. Some are related to electrical stability, some are related to safety.

If your turbine is small < 100KW, you are not likely to affect the distribution system stability. This is one less problem to worry about. The safety depends on your local utility and electrical code. The basic principle is that you shouldn't keep the utility line live when they turn their breaker off for maintenance. If you do, you could injure their personnel doing maintenance on the network. Look for Islanding electrical network problem for more details.

Finally, to answer your question, you need I x Xl volts above the line voltage. Where Xl is the complex impedance of the utility feeder at the point of connection and I is the current you want to supply. This math is vectorial. You have to worry about real and reactive currents. Power factor and harmonics can also come to haunt you.

Since the utility's impedance is very low, most systems use a current or power controlled scheme. You usually need to implement fairly complex controls.

Example: A 100KW 660V windmill should be connected to a utility feeder of at least 1MW to avoid control difficulties. Most feeder have an impedance around 6%. This means that to transfer your 100KW, you will need approximately 100KW/1MW*6%*660V = 3.9V above the utility (660V + 3.9V = 663.9V).

It is actually much more complex than this but the real answer will be near this value.

Read about generator's AVR, governors etc.. for details.

Have fun, contact a professional for the work.

Marcot, you said "...This is a problem that seasoned engineers haven completely solved yet..." Really? I guess all those windfarms I've seen in my travels are simply feathering as we engineers try to "completely solve" the problems. Actually many of those "problems" are created by well intentioned, but underknowledged, people such as the OP (who at least has the courtesy to acknowledge it). I agree with everything else you said though, it just seemed a bit unfair to state that "...seasoned engineers haven't completely solved..." them yet.

The problem is not completely solved yet. The stability and production capacity variations caused by windmills are only acceptable as long as they represent a small fraction of the total production on the network. A grid is not sustainable with a large fraction of wind or solar power (unless you accept black outs).

The control equations become very complex when multiple sources of variable production are installed "randomly" on the networks. While having production near the consumer help, the intermittent production generates many other problems. How do you coordinate the production of hundreds / thousands of windmills? You don't. You throttle the other power plants to keep the power balance about right. This affects voltage regulation and plant efficiency. The "free energy" transformed by a wind farm is often partially wasted by having to run the power plant nearby at low capacity. The power plant cannot be turned off because the wind might die down and more production will be needed rapidely.

Even if we install enough solar / wind fields to cover half our needs, we will still need full capacity from other sources to take the slack on those calm evenings. Meaningful storage is not available yet.

If you have a complete solution that is universal, you are unique and worth a lot of money. Bring it up!

I agree with most of your post, but disagree with where the problem lies. What I meant was that there are controllers that are capable of properly feeding the grid. The problem of accepting local generation lies with utilities who must ensure the sanctity of supply to all their customers. Utilities cannot dictate to customers the nature of their loads, but they do have some capability when it comes to generation, unfortunately it is not engineering in nature but legislative.

A prime example is that utilities now have to accept customer generation that is backfeeding the grid. In the early stages of development of alternative energy sources customers were flatly prohibited from connecting any form of generation to their meter and many meters had mechanisms that would prevent and/or flag reverse power through them. Ultimately there were laws passed that gave consumers the right to backfeed but only if they followed a strict set of guidelines that clearly protected the health and safety of both the distribution system and the workers maintaining it.

To your point though, there are many technologies that can be used for load-leveling in the face of rapidly varying, random generation such as wind and photovoltaic. In the 70's the Electric Power Research Institute (EPRI) invested heavily in the Battery Energy Storage Test (BEST) Facility in NJ, there were also investigations into pressurizing underground caverns with with reversible turbine-generators. Engineering concerns aside there remains the question of who pays for these huge facilities. If a utility owns the windmills then they may own the problem, but if an independent generation company owns them who pays for the leveling equipment, the customers, the investors, etc.? Perhaps the solution will be to force the problem back to the alternative energy suppliers who will have to provide a means (batteries and inverters seem most likely) to level the load before it gets to the grid. These answers will only come from laws passed that help determine who picks up the bill, and probably only after an investigation into some future blackout caused by the instability of said random generation. "It is a problem of control, not a control problem".

Many thanks for your help with this, I appreciate the advice and the references are exactly what I was looking for.